Little theory on applying something on a single vs multiple sources

Hey there,

When we are talking about adding glue to the whole mix, then send a stereo sum to a stereo comp and wield a tiny bit of compression. Now, I've wondered, is that really a glue effect between those tracks or do we simply benefit from a more controlled dynamic approach over all the sources equally? I mean, wouldn't it be the same if that stereo comp was in mono at the end of every track on a console at the same level? Or does a comp really somehow glue those transience together? I would just like to learn a bit more about the analog theory to not always refer to it as glue and actually knowing a little more than that.

Same for the stereo EQ, like a pultec, wouldn't it be the same if there was a pultec at the end of every channel because the level, after balancing, it will adjust itself accordingly anyways? No matter whether it was applied on a much higher level during equalizing the tracks? If we happened to do that on every channel separately?

I'd hugely appreciate it if someone could elaborate on that matter.

Mike

I'll start with some theory, in practice maybe some more experienced guy can answer better than me.

With a linear circuit you can process multiple sources and sum them and the result will be the same as sum them and process them.

I'll quote my self in something I post a couple of days ago talking about compressors

joaquins said:

Superposition theorem says that you can sum any sources and the result of the circuit will be the sum of the results of the solution with all the independent sources... But this is only with linear circuits, so the solution won't be the case for this because there is a cuadratic solution, same for rectifiers, true RMS, etc. So can't use this application and see the complex wave as fourier said just because this are not linear stuff, even if it's a constant wave but with harmonics, for this an square wave is as complex as a snare or any other musical content.

JS

Click to expand...

Compressors aren't linear, so won't be the same process the sum of signal or sum the signal processed. Same for the distortion added by the EQ or anything else, but assuming no distortion eq would be the same before and after, not compression. Delay and verb should be the same but depend the proccesor, isn't the same for gated reverbs for example... or any somehow level dependent process.

JS

If all sources are panned dead center, result might be the same. Imagine mixing only two instruments/channels, one instrument panned left side, the other instrument panned right side. A mono compressor per instrument will compress level above threshold by whatever ratio and if this only occurs to one of these instruments/sides, stereo image will shift to the other side. A stereo/bus compressor will compress level above threshold by whatever ratio, but operating both sides L/R the same, no matter which sides peaking signal causes compression, so keeping stereo image intact.

joaquins said:

I'll start with some theory, in practice maybe some more experienced guy can answer better than me.

With a linear circuit you can process multiple sources and sum them and the result will be the same as sum them and process them.

I'll quote my self in something I post a couple of days ago talking about compressors

joaquins said:

Superposition theorem says that you can sum any sources and the result of the circuit will be the sum of the results of the solution with all the independent sources... But this is only with linear circuits, so the solution won't be the case for this because there is a cuadratic solution, same for rectifiers, true RMS, etc. So can't use this application and see the complex wave as fourier said just because this are not linear stuff, even if it's a constant wave but with harmonics, for this an square wave is as complex as a snare or any other musical content.

JS

Click to expand...

Compressors aren't linear, so won't be the same process the sum of signal or sum the signal processed. Same for the distortion added by the EQ or anything else, but assuming no distortion eq would be the same before and after, not compression. Delay and verb should be the same but depend the proccesor, isn't the same for gated reverbs for example... or any somehow level dependent process.

JS

Click to expand...

so let's say I'd pass an eq at a higher level, single source, without causing distortion and later on lowering the balance and therefore also the amount that was applied, it'd be the same as having it sit within a sum of objects, applied by the same amount, i mean the applied eq boost remains the same, no matter whether given to every single source at a higher level or at a lower level on the whole stereo bus?

Harpo said:

If all sources are panned dead center, result might be the same. Imagine mixing only two instruments/channels, one instrument panned left side, the other instrument panned right side. A mono compressor per instrument will compress level above threshold by whatever ratio and if this only occurs to one of these instruments/sides, stereo image will shift to the other side. A stereo/bus compressor will compress level above threshold by whatever ratio, but operating both sides L/R the same, no matter which sides peaking signal causes compression, so keeping stereo image intact.

Click to expand...

Alright, I get it. But let's say i'd make sure every source would hit at the same level as it would be in a sum, panned left and right, the results should be quite similar right, same for an eq?
Whereas for the EQ, even though it'd pass the unit at a higher level that's not distorting, having the track later balanced and lowered in volume would act the same as having the signal in a sum of several signals at a lower level right? I forgot you call that effect in theory.

I don't understand that last thing about eq, why you mess with levels...

If no distortion or any other non linear stuff you could process in any order... Eq, level change (not compressing or clipping) and delay I guess (only delay I'm guessing) are linear process.

So for superposition it will be the same to process the sum of sources or sum the sources already processed... so you can boost, eq, atten then sum and will be the same as just sum and eq (same eq, boost and atten same amount)

I'm not agree with Harpo, if you have a mix in mono with a snare all way up and only it's signal get over threshold it will compress and get down the level of it self (snare) and all other stuff in the mix... If you compress before you mix the snare will hit the threshold and compress it self but the level of all other the mix will remain the same... so the result will be higher mix level for the same snare level. (ussing as example snare much higher than the rest of the signals)

JS

atticmike said:

Hey there,

When we are talking about adding glue to the whole mix, then send a stereo sum to a stereo comp and wield a tiny bit of compression. Now, I've wondered, is that really a glue effect

Click to expand...

Glue effect has been invented by marketing guys and trendy SE's to make you buy their latest gear or to make you believe they have a superior perception that you will never have because you haven't been chosen by the mighty gods.

between those tracks or do we simply benefit from a more controlled dynamic approach over all the sources equally?

Click to expand...

As others have already mentioned, you would have to apply exactly the same amount of instant Gain Reduction on all channels to achieve the same as one compressor on the bus. So, yes, bus compression creates something that cannot be done another way.

Same for the stereo EQ, like a pultec, wouldn't it be the same if there was a pultec at the end of every channel because the level, after balancing, it will adjust itself accordingly anyways? No matter whether it was applied on a much higher level during equalizing the tracks? If we happened to do that on every channel separately?

Click to expand...

You could replace a bus EQ with the same EQ on individual channels, it would produce the same effect, except if the characteristics of the EQ are somewhat non-linear in an audible way.